What's new in Summer 2021

Deepdesk Workflow Integrations, Deepdesk JavaScript SDK, and improvements on infra, autocompleter and more...

Less manual work, so you can focus on the human in the interaction.

Our new Workflow Integrations allow you to transform valuable information into fast actionable customer interactions.

Smarter conversations, with less manual work

Systems with content relevant to a conversation can be connected and made available for the agent to use directly in the chat.

Actionable rich content

Buttons make it easy for customers to respond, and allow information to be sent back to your systems for automated processing.

Increase conversion, reduce handling times, increase NPS

Ease of use drives promotion and conversion of your NBAs, and reduces manual work to process data. Easier and faster interactions make happy customers.

Content in your systems

All businesses use tools to maximize customer lifetime value, streamline services or manage customer data.
These CRMs, knowledge bases or next-best-action decisioning systems all contain valuable information about the customer. Information that is often relevant to the conversation.

With Workflow Integrations this content becomes available and actionable, directly in the conversation.

Make it actionable

Display a promotional message, with a call to action to self-service fulfillment. Confirm an email address. Upgrade to premium. Activate or reset accounts. Pay invoices. You name it.

All use cases that require the agent and visitor to exchange information and process that information in systems. There are so many use cases where a slightly smarter interaction could make a huge difference.

Workflow Integrations make it easier to exchange this information and automatically process that.

For automated processing

Manual work to process customer input into your systems takes up quite some time. Most of this information could - and should - be processed automatically. Speeding up the conversations, and freeing your agent to focus on the human side of the exchange.

How it works

We connect and configure your source

Knowledge bases, FAQs, CRMs, or tools like Pega are connected with our Workflow Integrations API. Your service can now propose a message, for a specific conversation.

The agent is prompted with the message

In your customer engagement platform, within the context of a specific conversation, the agent is now prompted with the proposed message. It is presented as it will be shown to the customer. So the agent knows exactly what he is about to send.

The customer receives and responds easily

The rich content is presented to the customer for easy and quick interacting. No need to type, and less reading for the agent.

Answers are processed automatically

When the customer hits one of the quick reply buttons, this is visible in the conversation, but also forwarded to your system for automated processing.

Increase conversion

Especially for next-best-actions, removing friction of actually sending the NBAs already increases the number actions promoted. Combined with the ease of use for customers to respond, this overall increases the conversion of your NBAs.

Reduce handling times

Buttons not only make it easier and faster for the customer to respond, but also make the response more compact and readable for the agent. More importantly, directly updating your system, removes manual work for the agent. Obviously reducing handling times, and human error.

Increase customer satisfaction

Customers like the ease of the buttons, instead of having to manually write a response. And with the reduced handling times the overall satisfaction of the interaction is increased, resulting in a higher NPS.

That's it for Workflow Integrations . But wait... There's more!

NEW: Deepdesk JavaScript SDK

With customer service platforms knocking on our door wanting to integrate Deepdesk themselves, we now have developed the Deepdesk JavaScript SDK .

The SDK allows platforms to integrate the Deepdesk Autocomplete overlay and widget, and style them as they wish. A first implementation was carried out by Tracebuzz with great success. Thanks to them a few minor issues were fixed and the SDK documentation has been improved.

Let us know if you want a key!


Regional infrastructure

With customers in various locations all over the world we needed our infrastructure to run close by, to make sure that our suggestions are offered real-time.
While modifying our infrastructure provisioning to be deployed where we want, we also updated and improved our security.

Autocompleter improvement: autocompleting variables

The autocompleter is better in autocompleting variables such as: parts of the day, or agent and customer names. When typing (part of) a name the autocompleter no longer stops, but continues to propose autocompletions.


We have localized our front-ends, including the tutorial, personal collection, and the demo app. Based on browser or user settings, the interface is shown in preferred locale.

Improved autoscaling

Autoscaling helps us to dynamically scale our infrastructure, and to provision resources matching the current workload. We have implemented a series of improvements that help scale faster, more efficiently, and reduce cost. Ronald our DevOps engineer wrote how we did that.


Continue reading in our spring post 🌱

Stay up to date

Header image by Etienne Girardet on Unsplash